Results
Study sample – a total of 74 survivors, almost all of them white Caucasian, formed the study sample. These subjects (40 males and 44 with standard risk [SR] ALL) were 13.5 – 38.3, median 21.1 years of age at the time of participation and 10.2 – 26.3, median 15.1 years from diagnosis. High risk (HR) disease was defined by any one of the following – age > 10 years, WBC > 50x199 /L, T cell phenotype, CNS involvement, mediastinal mass and t(9;22) translocation. The majority of the subjects (n=51, 69%) had received cranial irradiation, reflecting the era in which they were treated. All of the subjects underwent DXA and all but two had pQCT examination.
Total and trabecular BMC and trabecular vBMD were much lower at the (non-weight bearing) radial metaphysis than at the tibial metaphysis (Table 1), as in healthy subjects (Supplementary Tables 1 and 2) with which these values are closely similar. Cortical BMC and vBMD at the diaphyseal sites in both bones (Table 1) are also similar to published values (Supplementary Tables 1 and 2). Data from healthy subjects for cortical thickness are limited, with only one publication each at the 20% radius19 and 38% tibia,21 and two at the 66% tibia (Supplementary Tables 1 and 2).16,23 The mean values for this metric in the study sample were approximately the same as these published data; 2.33 v 2.7, 4.95 v 5.0 and 3.79 v 4.1 mm at the 20% radius, 38% tibia and 66% tibia respectively. Published data for polar SSI are even more limited, with none at the 38% tibia. Polar SSI values at the 20% radius and 66% tibia in the study sample (Table 1) are quite similar to the published values (Supplementary Tables 1 and 2).15,18,22 Polar SSI, a measure of bone strength (resistance to bending and twisting), and combining geometry and density, was progressively higher in the diaphyseal shafts with increasing distance from the distal metaphyses at the wrist and ankle (Table 1). The Z scores provided by comparison with the reference population are given in Table 2 and reveal a different picture of the health of both metaphyseal and diaphyseal bone.
The metrics from post-processed images are displayed in Supplementary Table 3. Striking and highly significant differences were evident between males and females at the metaphyseal sites (Table 3). These differences were seen in the non-weight bearing radius but not in the weight bearing tibia, although the directionality was the same; males had greater BMC and vBMD with improved architecture – smaller average and maximum hole sizes with fewer free ends and isolated points, and so higher connectivity. Likewise, apparent trabecular thickness and number were higher in males.
A different profile was seen at all of the diaphyseal sites (Table 3). Males had significantly greater total and cortical BMC but females had higher total vBMD, though this was not statistically significant, and there were significantly higher values in cortical vBMD for females, probably reflecting their lower cortical thickness. Most importantly, the polar SSI (a measure of bone strength) was significantly greater in males than in females at all diaphyseal sites.
There were no significant differences in any of the metrics at metaphyseal sites related to age at examination. The only metric which revealed a statistically significant difference at the diaphyseal sites was cortical vBMD – at all three sites. The subjects were divided into those 13-20 (N=30) and those > 20 years of age (N=42) and the younger group had lower values (20% radius 1141.3 [55.7] vs 1218.3 [39.8], p<0.0001; 38% tibia 1149.7 [48.4] vs 1203.5 [22.2]. p<0.0001; 66% tibia 1089.5 [44.5] vs 1129.2 [26.6], p<0.0001. This may relate to a smaller proportion of the younger subjects having attained peak bone mass.
With respect to ALL risk category, which is confounded by age at diagnosis – children 10 years of age and older are automatically high risk – there were no significant differences in BMC or vBMD, total or trabecular, at the metaphyseal sites between subjects who were SR and those HR at diagnosis, with the exception of Trabecular vBMD at the 4% radius, but there were notable differences in apparent bone architecture (Table 4). The SR group had smaller average and maximum hole sizes with fewer free ends and isolated points and so higher connectivity. Likewise, at both the 4% radius and 4% tibia the trabecular number was higher in SR than in HR subjects, reflecting more compact bone. At the diaphyseal sites there were statistically significant differences only with cortical vBMD; the HR subjects (N=29) having slightly higher values than those who had SR disease (20% radius 1204.7 [52.8] vs 1173.8 [62.5], p=0.032; 38% tibia 1193.6 [43.9] vs 1172.7 [48.1], p=0.048; 66% tibia 1125.4 [28.0] vs 1104.4 [44.7], p=0.028), with no differences in cortical thickness.
There were no statistically significant differences in any of the metrics at any of the sites between subjects with (N=27) and those without (N=45) a history of fracture (two missing data), nor between those who had received a bisphosphonate (N=14) and those who had not (N=58); two missing data. No differences were observed between those who had (N=49) and those who had not (N=23) received cranial irradiation (two missing data) with respect to the metaphyseal bone metrics. At the diaphyseal sites there were no consistent differences between irradiated and non-irradiated subjects.